Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 2450 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Node Classification | ogbn-arxiv | 15 | Mar 18, 2026 | ||
| Object Concealment | COCO subset | 15 | Mar 18, 2026 | ||
| Object Concealment | ImageNet subset | 15 | Mar 18, 2026 | ||
| Time-series Classification | StandWalkJump (SWJ) (test) | 15 | Mar 24, 2026 | ||
| Time-series Classification | AtrialFibrillation (AF) (test) | 15 | Mar 24, 2026 | ||
| Cancer subtype classification | TCGA | 15 | Mar 18, 2026 | ||
| Quadruped Locomotion | Slippery Slope real-world evaluation | 15 | Mar 18, 2026 | ||
| Expert-level Reasoning | Humanity's Last Exam 2,158 text-only | 15 | Mar 18, 2026 | ||
| Time Prediction | time tasks | 15 | Mar 18, 2026 | ||
| Extreme Weather Detection | Location answer-based questions | 15 | Mar 18, 2026 | ||
| Location Prediction | Location answer-based questions | 15 | Mar 18, 2026 | ||
| Numerical tasks | ZEPHYRUSBENCH (Entire dataset) | 15 | Mar 18, 2026 | ||
| Question Answering | General questions | 15 | Mar 18, 2026 | ||
| Web Agent Task Success | WebArena Overall 684 tasks | 15 | Mar 18, 2026 | ||
| Web Agent Task Success | WebArena Multi 29 tasks | 15 | Mar 18, 2026 | ||
| Web Agent Task Success | WebArena Reddit 106 tasks | 15 | Mar 18, 2026 | ||
| Web Agent Task Success | WebArena Gitlab 180 tasks | 15 | Mar 18, 2026 | ||
| Web Agent Task Success | WebArena 182 tasks (Admin) | 15 | Mar 18, 2026 | ||
| Web Agent Task Success | WebArena Shopping 187 tasks | 15 | Mar 18, 2026 | ||
| Semantic Segmentation | ADE20K (train) | 15 | Mar 17, 2026 | ||
| Fake News Detection | FakeTT | 15 | Mar 17, 2026 | ||
| Fake News Detection | FakeSV | 15 | Mar 17, 2026 | ||
| Adversarial Attack Transferability | ImageNet Target: IncRes 5000 images (val) | 15 | Mar 17, 2026 | ||
| Adversarial Attack Transferability | ImageNet Target IncV4 5000 images (val) | 15 | Mar 17, 2026 | ||
| Adversarial Attack Transferability | ImageNet Target: Visformer-S 5000 images (val) | 15 | Mar 17, 2026 |